We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 - 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!
Prior to this pandemic year of 2021, the term “supply chain” didn’t raise many red flags for most consumers, frankly because they didn’t have to think about it. Everything just happened. Buyers were so accustomed to getting things on schedule that it rarely became a regular topic of conversation.
That all changed in the second half of 2021. With the pandemic slowing down production lines and transportation in faraway places, the term “supply chain” is now regularly in headlines. This has been the greatest shock to global supply chains in modern history. Buyers often have to wait months for raw materials, durable goods, building materials, electronic devices, apparel, toys, and numerous other items. At the end of the calendar year, this remains a nagging problem that may continue well into 2022 – or even 2023.
As a result, supply-chain managers now are placing bets that may determine, in large measure, the fate of their companies. They’re desperate for visibility into all links in the chain – using portals many have never had before – but a number of them are flying blind with little or no control over the flow of their goods.
Supply chain managers struggle with how to best view and control logistics to get goods trapped in millions of 40×8-foot containers on ships waiting off ports in Oakland, Los Angeles, Long Beach, the eastern seaboard, and the Suez Canal onto trucks and trains and out to retailers. Solutions for this include those from companies such as SAP, Cin7, Oracle NetSuite, InfoPlus, and Anvyl. These suppliers make complex collections of point products that include controls for demand forecasting and management of import/export, inventory, shipping, suppliers, transportation, and warehousing.
These mostly legacy applications can be difficult to use, and they weren’t designed with optimal usability in mind. The good news is that there’s innovation happening in this market.
Welcome unified data analytics
Relative newcomer Incorta, which makes a software-as-a-service (SaaS)-based unified data analytics platform that includes the above functions, comes at the supply chain from a different perspective. Its single-screen platform puts all a company’s data into a single system, replacing various separate tools, to move data from source locations into a form that both line-of-business staff members and data scientists can use more effectively. This is the data analysis that’s used to project and/or identify supply-chain snags and find ways to solve them, similar to the way GPS routes drivers around traffic snarls.
“Incorta builds and deploys machine-learning models,” CIO Brian Keare told VentureBeat. “That’s in great part because our unified data analytics platform enables you to directly analyze raw, untransformed data – which is exactly the kind of data that is needed for machine learning.
“This has several implications: to begin with, it brings business analytics and data science together into closer alignment because both can operate on the same platform and work with the exact same data. For data scientists, that means no more developing models in a bubble and no more running with unrealistic data sets that ultimately fail in production. What’s more, it means that data scientists don’t have to spend so much time acquiring data, building a pipeline, and transferring results to another system to visualize and share results – a lot of this ‘grunt work’ is handled by the platform.”
“You can take a look at what your alternatives are, given that you’re out of stock on certain things,” Keare said. “Say your goods are stuck in on ships waiting to clear customs off the Port of Los Angeles. As opposed to a bunch of manual spreadsheets and trying to figure it out manually, you’ve got just one pane-of-glass view of what’s going on, and you can really look at what your alternatives are.”
How the AI is implemented
In order for technologists, data architects, and software developers to learn more about how to use AI, VentureBeat asked Keare the product applies AI.
VentureBeat: What AI and ML tools are you using specifically?
Brian Keare: Incorta’s unified data analytics platform bundles in and tightly integrates Spark, so that any library – whether open-source or commercial – can be used with it. Incorta also ships out of the box with favorites such as Scikit-learn, Spark-ML, FBprophet (Facebook prophet), and others. We also have utility libraries that make it easy to retrieve data from Incorta, save data frames back to Incorta, and then examine and visualize intermediate results in our embedded notebook interface.
VentureBeat: Are you using models and algorithms out of a box — for example, from DataRobot or other sources?
Keare: Incorta is a platform for analytics and developing ML models. We include some common Python libraries, but we also have a data API that can be used with external notebooks and third-party ML tools like DataRobot, Dataiku, H2O, and others. As a unified platform, Incorta is built with open standards and easily integrates with cloud-friendly tools and platforms.
VentureBeat: What cloud service are you using mainly?
Keare: Incorta is currently hosted on the Google Cloud Platform by default and other cloud services on request. We’re working with the other cloud vendors to make their platforms a turnkey, user-selectable option.
VentureBeat: Are you using a lot of the AI workflow tools that come with that cloud?
Keare: Incorta provides data scientists with an end-to-end platform that can be used to build AI workflows. More specifically, Incorta ingests data and then provides that data to a Spark-based ML/AI workflow. You can easily incorporate external tools as well, including other AI tools in the cloud.
VentureBeat: How much do you do yourselves?
Keare: Incorta takes data that is ingested into our UDAP (universal data access protocol) and makes it available to the ML/AI workflows developed within the Incorta product.
VentureBeat: How are you labeling data for the ML and AI workflows?
Keare: The ML workflow, including data labeling, is typically done using notebooks, which are then executed within the platform. These notebooks can be scheduled and orchestrated against other platform operations like data extraction.
VentureBeat: Can you give us a ballpark estimate on how much data you are processing?
Keare: While we don’t specifically track and measure how much data Incorta processes, we’re speaking with customers all the time and there are many who are processing many billions of rows of data every day using our platform. Some of the most valuable companies in the world today are Incorta customers.
Incorta is used by companies such as Starbucks, Broadcom, Duluth Trading Co., and Shutterfly.
Welcome to the VentureBeat community!
DataDecisionMakers is where experts, including the technical people doing data work, can share data-related insights and innovation.
If you want to read about cutting-edge ideas and up-to-date information, best practices, and the future of data and data tech, join us at DataDecisionMakers.
You might even consider contributing an article of your own!